Proceedings of the Annual Meeting of the Cognitive Science Society

Exemplar models (e.g., Nosofsky 1986) provide a highly influential account of the psychology of human category learning. However, the explanatory power of exemplar models may falter when applied to behavior outside of standard laboratory paradigms (Murphy, 2016) or even within the realm of traditional category learning experiments (Conaway & Kurtz, 2016; Kurtz & Wetzel, 2021). The present research poses new challenges that test the exemplar view within its wheelhouse of artificial classification learning tasks. Learners acquired categories based on two concentric circles (inner and outer) in feature space. Similarity-matched generalization tests reveal underlying global versus item-based category representation. Implications for exemplar and abstractive formal models of category learning are discussed.